feat(ui): add cache hit/miss filter to Request Logs

litellm/proxy/spend_tracking/spend_management_endpoints.py

@@ -2254,6 +2254,10 @@ async def ui_view_spend_logs(
     status_filter: str | None = fastapi.Query(
         default=None, description="Filter logs by status (e.g., success, failure)"
     ),
+    cache_hit_filter: str | None = fastapi.Query(
+        default=None,
+        description="Filter logs by cache state: 'hit' or 'miss'. Miss includes legacy rows with a null/unknown cache state",
+    ),
     model: str | None = fastapi.Query(default=None, description="Filter logs by model"),
     model_id: str | None = fastapi.Query(
         default=None,
@@ -2330,6 +2334,13 @@ async def ui_view_spend_logs(
             param="sort_order",
             code=status.HTTP_400_BAD_REQUEST,
         )
+    if isinstance(cache_hit_filter, str) and cache_hit_filter not in {"hit", "miss"}:
+        raise ProxyException(
+            message=f"Invalid cache_hit_filter: {cache_hit_filter}. Must be one of: hit, miss",
+            type="bad_request",
+            param="cache_hit_filter",
+            code=status.HTTP_400_BAD_REQUEST,
+        )
 
     try:
         is_admin_view: Final = _is_admin_view_safe(user_api_key_dict=user_api_key_dict)
@@ -2570,6 +2581,11 @@ def parse_date(date_str: str) -> datetime:
                 sql_params.append(status_filter)
                 p += 1
 
+        if cache_hit_filter == "hit":
+            sql_conditions.append("LOWER(cache_hit) = 'true'")
+        elif cache_hit_filter == "miss":
+            sql_conditions.append("(cache_hit IS NULL OR LOWER(cache_hit) != 'true')")
+
         if exclude_internal_health_checks:
             sql_conditions.append(f"api_key NOT IN (${p}, ${p + 1})")
             sql_params.extend(_INTERNAL_HEALTH_CHECK_API_KEYS)

tests/test_litellm/proxy/spend_tracking/test_spend_management_endpoints.py

@@ -106,6 +106,10 @@ def _iso(value):
             where["OR"] = where.get("OR", []) + [{"multi_team": True}]
         elif "status = 'success'" in cond:
             where["OR"] = where.get("OR", []) + [{"status": "success"}]
+        elif cond == "LOWER(cache_hit) = 'true'":
+            where["cache_hit"] = "hit"
+        elif cond == "(cache_hit IS NULL OR LOWER(cache_hit) != 'true')":
+            where["cache_hit"] = "miss"
         elif sess:
             where["session_id"] = {"contains": str(params[int(sess.group(1)) - 1]).strip("%")}
         elif status:
@@ -2444,6 +2448,96 @@ def filter_by_status(where):
         app.dependency_overrides.pop(ps.user_api_key_auth, None)
 
 
+@pytest.mark.asyncio
+async def test_ui_view_spend_logs_with_cache_hit_filter(client, monkeypatch):
+    base = {
+        "api_key": "sk-test-key",
+        "user": "test_user_1",
+        "team_id": "team1",
+        "spend": 0.05,
+        "startTime": datetime.datetime.now(timezone.utc).isoformat(),
+        "model": "gpt-4",
+        "status": "success",
+    }
+    mock_spend_logs = [
+        {**base, "id": "log1", "request_id": "req-hit", "cache_hit": "True"},
+        {**base, "id": "log2", "request_id": "req-miss", "cache_hit": "False"},
+        {**base, "id": "log3", "request_id": "req-legacy", "cache_hit": "None"},
+        {**base, "id": "log4", "request_id": "req-null", "cache_hit": None},
+    ]
+
+    def filter_by_cache(where):
+        cache_filter = where.get("cache_hit")
+        if cache_filter == "hit":
+            return [log for log in mock_spend_logs if str(log["cache_hit"]).lower() == "true"]
+        if cache_filter == "miss":
+            return [log for log in mock_spend_logs if str(log["cache_hit"]).lower() != "true"]
+        return mock_spend_logs
+
+    monkeypatch.setattr(
+        "litellm.proxy.proxy_server.prisma_client",
+        make_ui_spend_logs_mock_prisma(mock_spend_logs, filter_by_cache),
+    )
+
+    start_date, end_date = _default_date_range()
+
+    app.dependency_overrides[ps.user_api_key_auth] = lambda: UserAPIKeyAuth(
+        user_role=LitellmUserRoles.PROXY_ADMIN
+    )
+    try:
+        response = client.get(
+            "/spend/logs/ui",
+            params={
+                "cache_hit_filter": "hit",
+                "start_date": start_date,
+                "end_date": end_date,
+            },
+            headers={"Authorization": "Bearer sk-test"},
+        )
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total"] == 1
+        assert [row["request_id"] for row in data["data"]] == ["req-hit"]
+
+        response = client.get(
+            "/spend/logs/ui",
+            params={
+                "cache_hit_filter": "miss",
+                "start_date": start_date,
+                "end_date": end_date,
+            },
+            headers={"Authorization": "Bearer sk-test"},
+        )
+        assert response.status_code == 200
+        data = response.json()
+        assert data["total"] == 3
+        assert [row["request_id"] for row in data["data"]] == ["req-miss", "req-legacy", "req-null"]
+
+        response = client.get(
+            "/spend/logs/ui",
+            params={
+                "start_date": start_date,
+                "end_date": end_date,
+            },
+            headers={"Authorization": "Bearer sk-test"},
+        )
+        assert response.status_code == 200
+        assert response.json()["total"] == 4
+
+        response = client.get(
+            "/spend/logs/ui",
+            params={
+                "cache_hit_filter": "invalid",
+                "start_date": start_date,
+                "end_date": end_date,
+            },
+            headers={"Authorization": "Bearer sk-test"},
+        )
+        assert response.status_code == 400
+    finally:
+        app.dependency_overrides.pop(ps.user_api_key_auth, None)
+
+
 @pytest.mark.asyncio
 async def test_ui_view_spend_logs_with_model(client, monkeypatch):
     mock_spend_logs = [

ui/litellm-dashboard/src/components/networking.tsx

@@ -120,7 +120,7 @@
 })();
 export let proxyBaseUrl: string | null = _initialWorkerUrl ?? defaultProxyBaseUrl;
 if (isLocal != true) {
   console.log = function () {};
 }
 
 const getWindowLocation = () => {
@@ -235,7 +235,7 @@
   organization_id: string;
   organization_alias: string;
   budget_id: string;
   metadata: Record<string, any>;
   models: string[];
   spend: number;
   model_spend: Record<string, number>;
@@ -243,16 +243,16 @@
   created_by: string;
   updated_at: string;
   updated_by: string;
   litellm_budget_table: any; // Simplified to any since we don't need the detailed structure
   teams: any[] | null;
   users: any[] | null;
   members: any[] | null;
   object_permission?: ObjectPermission | null;
 }
 
 export interface CredentialItem {
   credential_name: string;
   credential_values: any;
   credential_info: {
     custom_llm_provider?: string;
     description?: string;
@@ -333,7 +333,7 @@
 
 let lastErrorTime = 0;
 
 export const handleError = async (errorData: string | any) => {
   const currentTime = Date.now();
   if (currentTime - lastErrorTime > 60000) {
     // 60000 milliseconds = 60 seconds
@@ -657,7 +657,7 @@
 
 export const budgetCreateCall = async (
   accessToken: string,
   formValues: Record<string, any>, // Assuming formValues is an object
 ) => {
   try {
     const data = await apiClient.post(`/budget/new`, {
@@ -676,7 +676,7 @@
 
 export const budgetUpdateCall = async (
   accessToken: string,
   formValues: Record<string, any>, // Assuming formValues is an object
 ) => {
   try {
     const data = await apiClient.post(`/budget/update`, {
@@ -2002,6 +2002,7 @@
   user_id?: string;
   end_user?: string;
   status_filter?: string;
+  cache_hit_filter?: string;
   /** Filter by model name (e.g. "gpt-4") */
   model?: string;
   /** Filter by model ID (litellm model deployment id) */

ui/litellm-dashboard/src/components/view_logs/RequestLogsFilters.test.tsx

@@ -69,6 +69,7 @@ describe("RequestLogsFilters", () => {
     for (const label of [
       "Team ID",
       "Status",
+      "Cache",
       "Key Alias",
       "User ID",
       "End User",
@@ -259,4 +260,37 @@ describe("RequestLogsFilters", () => {
 
     expect(await screen.findByText(label)).toBeInTheDocument();
   });
+
+  it.each([
+    ["", "All Requests"],
+    ["hit", "Cache Hit"],
+    ["miss", "Cache Miss"],
+  ])("shows the human label on the Cache trigger for %s", async (cacheState, label) => {
+    renderFilters(cacheState === "" ? {} : { [LOG_FILTER_IDS.CACHE_STATUS]: cacheState });
+
+    expect(await screen.findByText(label)).toBeInTheDocument();
+  });
+
+  it.each([
+    ["Cache Hit", "hit"],
+    ["Cache Miss", "miss"],
+  ])("selecting %s sets the cache filter to %s", async (label, expected) => {
+    const user = userEvent.setup();
+    const { set } = renderFilters();
+
+    await user.click(await screen.findByText("All Requests"));
+    await user.click(await screen.findByRole("option", { name: label }));
+
+    expect(set).toHaveBeenCalledWith(LOG_FILTER_IDS.CACHE_STATUS, expected);
+  });
+
+  it("selecting All Requests clears the cache filter", async () => {
+    const user = userEvent.setup();
+    const { set } = renderFilters({ [LOG_FILTER_IDS.CACHE_STATUS]: "hit" });
+
+    await user.click(await screen.findByText("Cache Hit"));
+    await user.click(await screen.findByRole("option", { name: "All Requests" }));
+
+    expect(set).toHaveBeenCalledWith(LOG_FILTER_IDS.CACHE_STATUS, undefined);
+  });
 });

ui/litellm-dashboard/src/components/view_logs/RequestLogsFilters.tsx

@@ -31,6 +31,12 @@ const STATUS_FILTER_ITEMS = [
   { value: "success", label: "Success" },
   { value: "failure", label: "Failure" },
 ] as const;
+
+const CACHE_FILTER_ITEMS = [
+  { value: ALL_VALUE, label: "All Requests" },
+  { value: "hit", label: "Cache Hit" },
+  { value: "miss", label: "Cache Miss" },
+] as const;
 const PAGE_SIZE = 50;
 
 const asString = (value: unknown): string => (typeof value === "string" ? value : "");
@@ -328,6 +334,27 @@ export function RequestLogsFilters({ get, set, teams, logsWindow }: RequestLogsF
         </Select>
       </DataTableFilterField>
 
+      <DataTableFilterField label="Cache">
+        <Select
+          items={CACHE_FILTER_ITEMS}
+          value={valueOf(LOG_FILTER_IDS.CACHE_STATUS) === "" ? ALL_VALUE : valueOf(LOG_FILTER_IDS.CACHE_STATUS)}
+          onValueChange={(next) =>
+            set(LOG_FILTER_IDS.CACHE_STATUS, next === null || next === ALL_VALUE ? undefined : next)
+          }
+        >
+          <SelectTrigger className="w-full">
+            <SelectValue placeholder="All Requests" />
+          </SelectTrigger>
+          <SelectContent>
+            {CACHE_FILTER_ITEMS.map((item) => (
+              <SelectItem key={item.value} value={item.value}>
+                {item.label}
+              </SelectItem>
+            ))}
+          </SelectContent>
+        </Select>
+      </DataTableFilterField>
+
       <KeyAliasFilterField
         value={valueOf(LOG_FILTER_IDS.KEY_ALIAS)}
         onChange={setter(LOG_FILTER_IDS.KEY_ALIAS)}

ui/litellm-dashboard/src/components/view_logs/log_filter_logic.test.tsx

@@ -83,6 +83,8 @@ describe("useLogFilterLogic", () => {
       { id: LOG_FILTER_IDS.SESSION_ID, value: "sess-1", param: "session_id" },
       { id: LOG_FILTER_IDS.END_USER, value: "end-user-1", param: "end_user" },
       { id: LOG_FILTER_IDS.STATUS, value: "failure", param: "status_filter" },
+      { id: LOG_FILTER_IDS.CACHE_STATUS, value: "hit", param: "cache_hit_filter" },
+      { id: LOG_FILTER_IDS.CACHE_STATUS, value: "miss", param: "cache_hit_filter" },
       { id: LOG_FILTER_IDS.MODEL_ID, value: "model-uuid-1", param: "model_id" },
       { id: LOG_FILTER_IDS.PUBLIC_MODEL_OR_SEARCH_TOOL, value: "gpt-4o", param: "model" },
       { id: LOG_FILTER_IDS.KEY_ALIAS, value: "alias-1", param: "key_alias" },

ui/litellm-dashboard/src/components/view_logs/log_filter_logic.tsx

@@ -20,6 +20,7 @@ export interface PaginatedResponse {
 export const LOG_FILTER_IDS = {
   TEAM_ID: "team_id",
   STATUS: "status",
+  CACHE_STATUS: "cache_hit",
   KEY_ALIAS: "key_alias",
   END_USER: "end_user",
   ERROR_CODE: "error_code",
@@ -35,6 +36,7 @@ export const LOG_FILTER_IDS = {
 export const LOG_FILTER_LABELS: Record<string, string> = {
   [LOG_FILTER_IDS.TEAM_ID]: "Team ID",
   [LOG_FILTER_IDS.STATUS]: "Status",
+  [LOG_FILTER_IDS.CACHE_STATUS]: "Cache",
   [LOG_FILTER_IDS.KEY_ALIAS]: "Key Alias",
   [LOG_FILTER_IDS.USER_ID]: "User ID",
   [LOG_FILTER_IDS.END_USER]: "End User",
@@ -170,6 +172,7 @@ export function useLogFilterLogic({
           user_id: userIdFilter,
           end_user: getFilterValue(columnFilters, LOG_FILTER_IDS.END_USER),
           status_filter: getFilterValue(columnFilters, LOG_FILTER_IDS.STATUS),
+          cache_hit_filter: getFilterValue(columnFilters, LOG_FILTER_IDS.CACHE_STATUS),
           model_id: getFilterValue(columnFilters, LOG_FILTER_IDS.MODEL_ID),
           model: getFilterValue(columnFilters, LOG_FILTER_IDS.PUBLIC_MODEL_OR_SEARCH_TOOL),
           key_alias: getFilterValue(columnFilters, LOG_FILTER_IDS.KEY_ALIAS),